sign in sign up
<<
Home , Detritivore, Niche differentiation, Species richness

CHAPTER 8 Towards a perspective on and functioning

Bradley Cardinale, Emmett Duffy, Diane Srivastava, Michel Loreau, Matt Thomas, and Mark Emmerson

8.1 Introduction that they exert a disproportionate influence over One of the most common questions asked by food web dynamics. This idea has fueled much researchers across a variety of scientificdisciplinesis debate over the prevalence of omnivory in food ‘How does the number of nodes connected together webs (Polis and Strong 1996, Thompson et al. 2007, into a network influence the efficiency and reliability Yodzis 1984) and whether the increased number of of that network?’. While social scientists and epide- feeding links that result from omnivory increases or miologists might think of ‘nodes’ and ‘connections’ decreases the stability of energy flow through a as people interacting within a social network, com- food web (McCann et al. 1998, MacArthur 1955). puter scientists, neurologists, and civil engineers Identifying that represent influential nodes would instead think of servers connected together in has also been one of the primary goals in the search a world-wide web, synapses connecting neurons in for ‘ecosystem engineers’ (Jones et al. 1994), ‘key- the brain, or hubs connecting to other hubs in a stone species’ (Paine 1966, Power et al. 1996) or transportation network or telecommunications grid other types of ‘strong interactors’ (Wootton and (Albert and Barabasi 2002, Newman 2003). Regard- Emmerson 2005) that might have cascading effects less of the particular study system, all of these on the diversity and of species at a variety individuals ask similar questions about how the of different trophic levels (Paine 1966, Carpenter number of nodes and connections among nodes et al. 1987, Elser et al. 1988). influence the efficiency and reliability by which In the 1990s, ecologists began to pursue a slightly information, disease, energy, or matter is transmitted different perspective on food webs. This perspec- throughout that network. tive focused not on the cascading impacts of indi- Within the field of , one of the oldest and vidual species, but rather on how the number of most fundamental questions asked by researchers is species that comprise any single might ‘How does the number of species interacting within control fluxes of energy and matter. Research in this a food web influence the efficiency and reliability area was generally referred to as Biodiversity effects by which energy and matter are transmitted on Ecosystem Functioning (BEF for short), and was through that web?’. Research on this topic can be often justified on grounds that (1) loss of biological broadly divided into two foci. Historically, much diversity ranks among the most pronounced chan- attention in ecology has focused on identifying ges to the global environment (Sala et al. 2000, those taxa that are the most influential nodes in a Pimm et al. 1995), and (2) reductions in diversity, food web. For many years, it has been thought that and corresponding changes in species composition, some subset of species might represent ‘hubs’ of may alter fluxes of energy and matter that underlie interactions and/or exhibit such strong interactions important services that provide to

105 106 BIODIVERSITY, ECOSYSTEM FUNCTIONING, AND HUMAN WELLBEING

(a) 25 100 (b)

20 80

15 60

10 40 % Studies Number species 5 20

0 0 123456 Trophic level

Figure 8.1 (a) Summary of the biological complexity of biodiversity-ecosystem functioning (BEF) studies performed to date. On the x-axisisthe number of trophic levels included in different experiments. On the left-hand y-axis (plotted as grey bars) is the mean number of species per trophic level. On the right hand y-axis (plotted as triangles) is the percentage of studies that have included 1, 2, or more trophic levels. Note that 93 per cent of BEF experiments have focused on a single trophic level composed of a mean seven species. (b) An example of the complexity of a real, yet still relatively simple natural food web in a salt marsh (from Lafferety et al. 2007). Note that within this system there are dozens of species (nodes) and hundreds of feeding links (lines connecting nodes) among , , predators and parasites that span six or more trophic levels. Figure reproduced with permission from K. Lafferty.

humanity (e.g. production of food, pest/disease simplifications are justified, or alternatively, whether control, water purification, etc. Daily 1997, Chapin et they have led ecologists to potentially erroneous al. 1998). While the value of BEF research for conser- conclusions. However, what is clear is that a large vation biology and management has been questioned body of research in ecology has shown that interac- by some (Schwartz et al. 2000, Srivastava and Vellend tions of species across trophic levels can have cas- 2005), there is a more fundamental reason for the cading impacts that influence the diversity and recent prominence of this topic. BEF is one of the few biomass of organisms at numerous levels in a food research topics in ecology that examines how bio- web. At the very least, this suggests that the past focus logical variation per se acts as an independent variable of BEF on diversity within single trophic levels may to regulate key and ecosystem-level pro- be insufficient to quantitatively predict, and perhaps cesses (Naeem 2002b). Understanding the ecological even qualitatively reflect, the ecological consequences consequences of variation among species has shown of diversity loss. much potential to complement our historical focus on In this chapter, we continue with the development the ecological impacts of highly influential species. of an idea that originated with other authors who Although the BEF paradigm has evolved consid- have argued that, in order to understand how erably over the past 15–20 years and been increasingly alters the functioning of whole ecosys- applied to a variety of organisms and ecosystems, tems, ecologists will likely need to merge modern studies have continued to focus mostly on simplified paradigms of BEF with much more classic ideas in ‘model’ communities. In fact, the typical experiment food web ecology that consider not only the func- has manipulated an average of just seven species in an tional role of diversity within trophic levels, but the average of just one trophic group (Fig. 8.1(a)). Such interactions of species across trophic levels (Duffy minimal levels of complexity are far from the realities et al. 2007, Bruno and Cardinale 2008, Petchey et al. of natural food webs, where, even for some of the 2004a). Our chapter is organized as follows. In Sec- simplest communities, species interact within webs tion 8.2 we brieflyreviewfive hypotheses about how composed of hundreds of species spanning many fluxes of energy and matter through a food web trophic levels (Lafferty et al. 2006, Polis 1991, Martinez might depend on the diversity of species comprising 1992). At present, it is unclear whether such over- a web. Those hypotheses are divided into those that TOWARDS A FOOD WEB PERSPECTIVE ON BIODIVERSITY AND ECOSYSTEM FUNCTIONING 107 contrast diversity effects within different trophic impacts on processes that are disproportionate to levels versus those that focus on diversity effects their or biomass. Duffy’s(2002)paper across trophic levels. In Section 8.3 we outline the was one of the firsttocallforamergerofBEFand empirical support for or against these hypotheses, food-web theory, and the hypotheses put forth in emphasizing that most are still unresolved and in that paper were useful, in part, because they need of testing. In the final Section 8.4, we outline represented an alternative to those posed by a just a few of the areas of research that we believe number of other authors. For example, some have will be fruitful as ecologists move towards an inte- argued that extinction at higher trophic levels gration of BEF into food-web ecology. may, in fact, have less impact on ecological pro- cesses than extinction at lower trophic levels. These arguments have usually been based on the 8.2 Five early hypotheses about idea that are more generalized in their use multi-trophic biodiversity and ecosystem of resources than historically appreciated, either function because the extent of omnivory and intra- has been underestimated (Rosenheim et 8.2.1 Diversity effects within trophic levels al. 1995, Holt and Polis 1997, Polis and Holt 1992), 8.2.1.1 Top-down effects of diversity grow increasingly or because animals can ‘switch’ among different strong at higher trophic levels prey species by moving across (Polis et al. 1997, McCann et al. 2005). generalization has been proposed to dampen the effects of con- C1 Cn sumer diversity on prey populations (Finke and +++ Denno 2005, Snyder and Ives 2003).

H1 Hn 8.2.1.2 Increasing diversity of a resource reduces the strength of top-down control by consumers ++

P P 1 n C + – N

R1 Rn Early hypotheses proposed that species extinction from higher trophic levels was likely to have greater The majority of BEF studies performed to date have impacts on the functioning of ecosystems than taken a ‘top-down’ perspective, meaning that they extinction from lower trophic levels (Table 8.1). have examined how diversity within a given Duffy (2002) argued that three characteristics trophic level impacts the fraction of resources con- potentially make ecological processes more sensitive sumed, and production of biomass, by that focal to extinction by consumers than plants: (1) because trophic level. In contrast, diversity may also have species at higher trophic levels have lower popula- ‘bottom-up’ effects on the dynamics of food webs, tion sizes and are under stronger anthropogenic meaning that the diversity of resources may influ- pressure than most wild plants, higher trophic levels ence how efficiently those resources are consumed face greater risks of extinction and higher rates of and converted into biomass by higher trophic levels species loss; (2) assemblages have lower (Table 8.1). At least three hypotheses have been overall richness and higher degrees of resource proposed to explain how resource diversity might specialization, leading to less ‘functional redun- influence trophic dynamics: (2.1) the variance in dancy’ and limited potential for surviving species to edibility hypothesis argues that a more diverse prey compensate for processes performed by lost coun- assemblage is more likely to contain at least one terparts; and (3) unlike plants, consumers often have species that is resistant to consumers (Leibold 1989, Table 8.1 Five early hypotheses about multi-trophic BEF. What do the data say?

Section Hypothesis Key reference(s) Section Balance of evidence Certainty Key references

8.2.1 Diversity effects within trophic levels 8.2.1.1 Top-down effects of diversity grow increasingly Duffy (2002) 8.3.1.1 The balance of evidence is not consistent with this hypothesis. Medium Balvanera et al. (2006), strong at higher trophic levels. Recent meta-analyses have found no difference in the to high Cardinale et al. (2006a, direction or magnitude of diversity effects for groups of 2007) producers, herbivores, , or predators. In fact, there tends to be considerable generality such that decreases in decrease the efficiency of resource capture and the amount of biomass produced by any given trophic group. 8.2.1.2 Increasing diversity of resources reduces the Leibold (1989), 8.3.1.2 The balance of evidence is consistent with this hypothesis. Low to Andow (1991), Hillebrand strength of top-down control by consumers. Duffy (2002), Summaries suggest that consumption of lower by higher medium and Cardinale (2004) Ostfeld and trophic levels is reduced when a resource base is more LoGiudice diverse. Note, however, that most of the available data (2003), Root comes from studies that have not directly manipulated the (1973) richness of resources. Several controlled experiments have provided counter-examples, so the generality of this hypothesis remains unclear. 8.2.2 Diversity effects across trophic levels 8.2.2.1 Top-down effects of consumer diversity oppose Holt & Loreau 8.3.2.2 Recent meta-analyses suggest that the top-down effects of Low Srivastava et al. (in press) the bottom-up effects of resource diversity. (2002), Thébault consumer diversity are qualitatively different than the and Loreau bottom-up effects of resource diversity. However, these (2003, 2005) effects have not been opposing as suggested by this hypothesis. Note, however, that few studies have simultaneously manipulated the richness of species at adjacent trophic levels, so conclusions are tentative. 8.2.2.2 Diversity effects on biomass production and Holt and Loreau 8.3.2.1 The balance of evidence does not support this hypothesis. Of the Low Mulder et al. (1999), Duffy et resource capture by a given trophic level are (2002), Thébault few experiments that have manipulated species richness in the al. (2005), Wodjak (2005), reduced in the presence of a higher trophic and Loreau presence vs. absence of a higher trophic level, results are and this chapter level. (2003) decidedly mixed. Analyses presented in this chapter further show no evidence that the effects of diversity on plant biomass differ for experiments performed in the presence vs. absence of herbivores. 8.2.2.3 Trophic cascades are weaker in diverse Strong (1992) 8.3.2.3 Experiments and data summaries to date have been equivocal None Schmitz et al. (2000), Borer et communities. and contradictory. At present, there is no clear reason to al. (2005), Cardinale et al. accept or reject this hypothesis. (2003, 2006b), Wilby et al. (2005), Snyder et al. (2006), Finke and Denno (2005), Byrnes et al. (2006) TOWARDS A FOOD WEB PERSPECTIVE ON BIODIVERSITY AND ECOSYSTEM FUNCTIONING 109

Duffy 2002); (2.2) the dilution hypothesis (Ostfeld and tially more complex. While predator diversity LoGiudice 2003), which has also been called the generally decreases prey biomass, prey diversity resource concentration hypothesis in the agro-ecology can increase or decrease biomass depending on literature (Root 1973), suggests that specialist con- how different -history trade-offs influence the sumers become less efficient at finding and attacking coexistence of prey. their resource in a diverse prey assemblage; and (2.3) the balanced diet hypothesis suggests that a more diverse prey assemblage provides a more complete 8.2.2.2 Diversity effects on biomass production nutrition and, as a result, leads to higher consumer and resource capture by any focal trophic level biomass (DeMott 1998). While hypotheses (1) and (2) are reduced in the presence of higher trophic levels predict that trophic efficiency will decrease as the diversity of resources increases, (2.3) predicts the CЈ opposite.

C1 Cn 8.2.2 Diversity effects across trophic levels + (0) 8.2.2.1 Top-down effects of consumer diversity oppose R the bottom-up effects of resource diversity

In their recent review, Duffy et al. (2007) used the C1 Cn terms ‘horizontal’ and ‘vertical’ diversity to distin- – guish between the richness of species within a trophic level and the richness of trophic levels that R R 1 n comprise a food web. They argued that one of the primary limitations in merging BEF with food-web An important, but still unresolved issue is whe- theory is knowing how the impacts of divesity ther the overall impacts of diversity loss at within trophic levels depend on the length of food adjacent levels are opposing or reinforcing, chains (i.e. how horizontal and vertical diversity antagonistic or synergistic. Hypotheses (2.1) and interact). The first step in overcoming this limitation (2.2) suggest that consumer diversity tends to is to ask how the diversity effects of any single enhance the flux of resources from lower to trophic level are altered by the presence or absence higher trophic levels, whereas resource diversity of the next highest trophic level. Holt and Loreau tends to reduce these fluxes. Collectively, these (2002) used simple consumer–resource models to two hypotheses lead to a third hypothesis: that argue that the effects of plant diversity on extinction of species from adjacent trophic levels uptake and plant biomass production are reduced will have opposing impacts on the flux of energy in the presence of herbivores. This occurs because and matter through a food web (Table 8.1). This herbivory selects for by poor plant prediction has received some theoretical support competitors that are also the most tolerant to from mathematical models showing that simulta- consumption by herbivores. Subsequent models by neous changes in diversity from consumers and Thébault and Loreau (2003) also suggested that their resource leads to countervailing effects on addition of higher trophic levels might qualita- total resource use and biomass production (The- tively alter diversity–production relationships at bault and Loreau 2003, Thebault and Loreau 2005, lower levels; however, the direction of these Holt and Loreau 2002). Fox (2004b) provided a impacts depends on both the nature of trade-offs counter example in which he used Lotka–Volterra between a plant’s competitive ability and ability to models to show that the joint response of prey resist herbivory, and on the degree of consumer biomasstopreyandpredatordiversityispoten- specialization. 110 BIODIVERSITY, ECOSYSTEM FUNCTIONING, AND HUMAN WELLBEING

8.2.2.3 Trophic cascades are weaker in effects are stronger at higher trophic levels. diverse communities Balvanera et al. (2006) reviewed 103 studies in which they could examine 400þ correlation coef- ficients relating species richness to a variety of C Ј C Ј 1 n ecological processes. They found no evidence for – differential correlations between diversity and any – of the response variables at various trophic levels. C C 1 n Similarly, Cardinale et al. (2006a) collated data from 111 experiments that have manipulated spe- cies richness and examined how this aspect of R diversity impacts the capture of resources and production of biomass. Their analyses compared In his seminal critique of the empirical evidence for four trophic groups: (1) microalgal, macroalgal, or trophic cascades, Strong (1992) argued that cas- herbaceous plants assimilating or water, cades are ‘a relatively unusual sort of food web (2) protozoan or metazoan herbivores consuming mechanics . . . over the full range of ecological com- live algal or herbaceous plant tissue, (3) protozoan munities, evidence is that these cascades are or metazoan predators consuming live prey, and restricted to fairly low-diversity places where great (4) bacterial, fungal or metazoan detritivores con- influence can issue from one or a few species’.He suming dead organic matter. They showed that, on went on to suggest that trophic cascades are ‘all average, experimental reduction of species rich- wet’, meaning they occur primarily in aquatic eco- ness decreases the standing stock abundance or systems where communities are characterized by biomass of the focal trophic group, resulting in less linear, low-diversity food chains. In contrast, he complete resource use by that group (Fig. 8.2). argued that terrestrial food webs are more reticu- However, the standing stock of, and resource late and ‘consumption is so differentiated in spe- depletion by, the most diverse polycultures were ciose systems that its overall effects are buffered’. indistinguishable from those of species that per- The idea that diversity modifies the strength of formed best in monoculture. Importantly, the trophic cascades can be broken down into at least authors could not detect any statistical difference two distinct hypotheses: (1) increasing the diversity in the magnitude of diversity effects among the of species comprising secondary consumers C’ four trophic groups. tends to decrease the strength of indirect effects on Collectively, these meta-analyses suggest there a basal resource R, and (2) increasing diversity of is considerable generality in the way that the primary consumers C tends to decrease the indirect diversity of species impact resource capture and effects of C’ on R. This latter hypothesis is very biomassproductioninfoodwebs.Thefactthat much an extension of hypotheses (2.1) and (2.2), as Cardinale (2006a) and Balvanera (2006) both found all of these rely on the assumption that an that the BEF relationships did not change dra- increasing diversity of resources tends to reduce the matically across trophic levels could imply that, if top-down impacts of consumers on food-web niche complementarity is the main mechanism dynamics (Table 8.1). driving these patterns, then the degree of niche complementarity could be similar across trophic groups. Identifying whether the mechanisms that 8.3 What do the data say? dictate BEF relationships are the same across dif- 8.3.1 Diversity effects within trophic levels ferent levels of biological organization is a key next step in BEF research (a point we return to in 8.3.1.1 Are diversity effects stronger at higher Section 8.4.1). Although studies to date show trophic levels? (Hypothesis 2.1) considerable generality in diversity effects across Empirical evidence gathered to date does not trophic levels, we should emphasize that there still appear to support the hypothesis that diversity tend to be fewer absolute numbers of species at TOWARDS A FOOD WEB PERSPECTIVE ON BIODIVERSITY AND ECOSYSTEM FUNCTIONING 111

Producers Herbivores Predators Detritivores Producers Herbivores Predators Detritivores 3

2 4

1 2 0 LRmax (SST)

LRmean (SST) 0 –1

–2 –2 35, 15, 3 4, 9, 1 4, 1, 1 3, 3, 0 7, 21, 11 3, 11, 0 3, 0, 3 1, 2, 0 8 4 1

2 0

–1

0 LRmax (RD) LRmean (RD) –2 –2 5, 9, 0 0, 7, 1 1, 18, 0 16, 13, 0 1, 10, 1 1, 9, 0 0, 16, 0 3, 18, 0

Figure 8.2 Summary of the results of experiments that have manipulated the richness of species in four trophic groups t (producers, herbivores, predators, and detritivores), and examined how richness impacts the standing stock abundance or biomass of t (SST – top graphs) or the fraction of resources depleted by t (RD – bottom graphs). The y-axes in all graphs give the diversity ‘effect size’, measured using two log ratios. LRmean (left graphs) compares SST and RD in the most diverse polyculture used in a study to the average of all monocultures. LRmax (right graphs) compares SST and RD from the most diverse polyculture used in a study to the species having the highest values of SST or RD in monoculture. Each data point is the mean effect size for all replicates in an experiment ± 95 per cent CI. Dashed horizontal grey lines give the 95 per cent CI for all experiments combined based on results rom a mixed model ANOVA. Numbers below each figure are the number of studies that have shown significantly positive effects of diversity, no effect, or negative effects of diversity. Data are from Cardinale et al. (2006a).

higher trophic levels, and that these species tend to resource in a diverse prey assemblage, best be disproportionately prone to extinction (a point accounted for the observed patterns. A summary of we return to in Section 8.4.2). Thus, it is still rea- aquatic studies by Hillebrand and Cardinale (2004) sonable to hypothesize that food webs can tolerate tallied results from 172 experimental manipulations fewer at higher trophic levels before of herbivores and showed that consumption of ecosystem functioning is altered. algal biomass generally declined with increasing algal species richness. Although these patterns are 8.3.1.2 Does resource/prey diversity weaken consistent with hypothesis (2.2), some caution is the strength of top-down control? (Hypothesis 2.2) warranted when interpreting these summaries, Empirical evidence gathered thus far is mostly since the studies reviewed did not manipulate consistent with the hypothesis that increasing prey directly, and many potentially diversity tends to reduce the impacts of consumers confounding factors were not controlled for. This on prey. Andow (1991) tallied the results of 200þ caveat is particularly important when considering studies of herbivorous and found that the mixed results from the limited number of more than half of the species had lower experiments that have manipulated resource population sizes on plant polycultures as opposed diversity directly. Several studies do provide evi- to monocultures. He argued that the resource con- dence consistent with the variance-in-edibility centration hypothesis, in which specialist con- hypothesis (Steiner 2001, Duffy et al. 2005), or for the sumers have a more difficult time finding their dilution hypothesis (Keesing et al. 2006, Wilsey and 112 BIODIVERSITY, ECOSYSTEM FUNCTIONING, AND HUMAN WELLBEING

Aquatic Terrestrail Figure 8.3 A summary of the impact of plant species richness on the 2 production of plant biomass when herbivores are present or absent in experimental units. Data were taken from the summaries of Cardinale et al.

) (2006, 2007). The log ratio of plant biomass in the most diverse polyculture m

B Bp to biomass in the average monoculture Bm was analyzed using a mixed / 1 p model ANOVA with herbivores (y/n), ecosystem (aquatic vs. terrestrial), and B their two-way interaction included as fixed effects, experiment accounted for as a random effect, and observations weighted by the inverse of their 0 variance. Analyses indicate that the impacts of plant diversity on plant FPO F ¼ P production do not differ when herbivores are absent vs. present ( 0.01, 90 15 ¼ 0.92), and that this conclusion is consistent among ecosystems (F ¼ –1 9 0.07, P ¼ 0.80 for interaction). These data and analyses should not be 29 taken as conclusive evidence that herbivores do not impact plant on plant biomass, In( – Effect of plant species richness Effect diversity biomass relationships since the studies summarized here differ in –2 many ways that cannot be explicitly accounted for. However, these data can serve as a null hypothesis for experiments that explicitly manipulate plant Yes No Yes No diversity in the presence versus absence of higher trophic levels. Herbivores present?

Polley 2002, Otway et al. 2005) where increasing bottom-up effects of resource (detrital) diversity on diversity of resources leads to reduced consumption consumption of dead organic matter. Their meta- by higher trophic levels. Other studies provide analysis indicated that reductions in support for the balanced diet hypothesis, showing diversity generally led to reductions in rates of that mixed diets of primary producers tend to , but changes in the diversity of enhance herbivore growth and biomass accumula- detrital resources led to no detectable change in tion (Pfisterer et al. 2003, DeMott 1998). Thus, decomposition. The implication is that consumer, although the balance of evidence appears consistent but not resource diversity, impacts consumption with hypothesis (2.2), these conclusions should be and energy flow in ‘brown’ food webs (- considered tentative. consumer). However, an important point to keep in mind is that the resources studied by Srivastava et al. (2008) are ‘dead’, meaning they are non-living 8.3.2 Diversity effects across trophic levels resources that have no potential to show dynamic 8.3.2.1 Do top-down effects of diversity differ coupling to their consumers. A number of mathe- from bottom-up effects? (Hypothesis 2.1) matical models suggest that diversity–function To date, studies that have simultaneously manipu- relationships could be qualitatively different when lated the richness of species at adjacent trophic resources are ‘living’, such as in ‘green’ food webs levels are rare (Fig. 8.1(a)), and it is difficult to draw (i.e. plant-based systems) where populations have many general conclusions about the direction of the potential to respond to changes in the density of top-down versus bottom-up effects of diversity in their consumers (Loreau 2001, Ives et al. 2005). The food webs. However, a recent meta-analysis by potentially important contrast between systems that Srivastava et al. (2008) suggests that hypothesis (2.1) have dynamic (living) vs. non-dynamic (non-living) is is not supported in detrital systems. These authors an issue that we return to in Section 8.4.1. For now, compiled the results of 90 experiments reported in suffice it to say that we do not know whether the 28 studies of detritivores to ask ‘Do changes in results of Srivastava et al. (2008) are specifictodetrital consumer (i.e. detritivore) diversity have the same systems, or whether they hold more generally. effect on rates of resource consumption as changes in resource (i.e. detrial) diversity?’. To address this 8.3.2.3 Are diversity effects at one trophic level question, they compared the top-down effects of altered by higher levels? (Hypothesis 2.2) consumer (detritivore) diversity on the consump- Only a handful of experiments have manipulated tion of dead organic matter (decomposition) to the the richness of species in a focal trophic level and TOWARDS A FOOD WEB PERSPECTIVE ON BIODIVERSITY AND ECOSYSTEM FUNCTIONING 113 then simultaneously manipulated the presence/ ecosystems found no evidence that variation in the absence of a higher trophic level. Mulder et al. (1999) strength of cascades was related to the richness of varied plant diversity in the presence and absence of predators, herbivores, or plants (Borer et al. 2005). In insect herbivores in a grassland plant assemblage. In contrast, a limited number of experiments have the absence of herbivores, plant biomass increased manipulated the diversity of predators at top trophic with plant diversity, whereas when insects were levels and shown that diversity can indirectly alter present, they fed heavily on species with interme- plant biomass by changing rates of herbivory. Cas- diate biomass, weakening the impact of plant cading effects of predator diversity have been dem- diversity and biomass. Conversely, in a seagrass onstrated in agricultural (Cardinale et al. 2003, Wilby system, effects of herbivore richness on plant pro- et al. 2005, Snyder et al. 2006), salt marsh (Finke and duction were stronger in the presence of a higher Denno 2005), and systems (Byrnes et al. trophic level (crabs) than in their absence (Duffy et al. 2006), and have been attributed to non-additive 2005), which presumably occurred because of tra- interactions (Cardinale et al. 2003, Cardinale et al. deoffs between species abilities to compete for 2006b), omnivory (Bruno and O’Connor 2005), intra- resources versus resist predators. In other experi- guild predation (Finke and Denno 2005), and chan- ments, addition of a higher trophic level changed not ges in herbivore behavior (Finke and Denno 2005, only the magnitude but also the sign of the Byrnes et al. 2006). Yet, the magnitude and direction diversity–function relationship at the prey level (e.g. of predator richness impacts on plant biomass and Hattenschwiler and Gasser 2005, Wojdak 2005). production have been inconsistent among studies We have been able to further examine hypothesis (see Bruno and Cardinale 2008 for a review). Thus, (2.2) by collating data from the meta-analyses of although predator richness frequently has cascading Cardinale et al. (2006a, 2007) for studies that have impacts on food-web properties, it is difficult at this manipulated the richness of primary producers. We point in time to predict whether these cascading divided experiments into those that did versus did effects generally increase or decrease plant biomass. not allow herbivores access to experimental plots or Therefore, at present, there is no clear evidence that pots, and then compared how plant diversity influ- can be used to accept or reject Strong’s (1992) ence plant biomass between the two types of studies. hypothesis that trophic cascades are restricted to Although plant species richness generally increased low-diversity linear food chains. the production of plant biomass, we found no evi- dence that herbivores alter the magnitude of plant diversity effects (Figure 8.3). This was true for studies 8.4 Where do we go from here? performed in both aquatic as well as terrestrial eco- 8.4.1 Detailing mechanisms: niche partitioning systems. Although these analyses are far from con- and life-history tradeoffs clusive, when taken with the mixed results of experiments they suggest that widespread support William Dillard, founder and Chairman of Dillard’s for hypothesis (2.2) is presently lacking. department stores, once said that the three most important factors for the success of a business are 8.3.2.4 Are trophic cascades weaker in ‘location, location, location’. Similarly, we believe diverse communities? (Hypothesis 2.3) that the three most important factors that will Experiments and data summaries that have addres- determine the success of the BEF paradigm will be sed hypothesis (2.3) to date have been equivocal and our ability to identify mechanisms, mechanisms, contradictory. Schmitz et al. (2000) performed a mechanisms! Understanding the mechanisms that meta-analysis of 14 terrestrial experiments that underlie diversity effects essentially requires that manipulated higher predators and found evidence researchers return to several of ecology’s classic that the cascading effects of predator removal on questions about how niche partitioning and life- plant damage were weaker in systems that had history tradeoffs allow species to coexist. Chesson higher herbivore diversity. A more comprehensive (2000) provided what is perhaps the most elegantly analysis of trophic cascades measured in a variety of organized summary of the mechanisms that allow 114 BIODIVERSITY, ECOSYSTEM FUNCTIONING, AND HUMAN WELLBEING

d SK B = 1 + a(S – 1) a = 0 w B 0 < a < 1 d biomass, Community a = 1 Frequency of utilization Frequency w Species richness, S Resource dimension

Figure 8.4 (a) Solutions to Lotka–Volterra equations showing how species richness affects community biomass production for differing levels of interaction strength. Note there is a positive, but decelerating relationship between B and c for all 0 < a < 1. This is an inevitable consequence of niche packing (insets) where the addition of species to a system with finite resource forces the average species to occupy a smaller fraction of resource space. Thus, the more species there are, the less each species contributes to resource capture and biomass production, on average.

coexistence. He showed that, for a wide variety of systems (Loreau 2004, Tilman et al. 1997c, Ives et al. mathematical models, coexistence is ultimately 2005, Cardinale et al. 2004). The discrete time determined by the balance of two interacting forces, Lotka–Volterra models of competition serve as an which he called equalizing and stabilizing. Equalizing example (Cardinale et al. 2004), where the biomass of forces are those that minimize differences in the fit- any species i in a local community can be described as ness of species, causing interspecific interactions to 2 0 13 have weaker influence over . PN b ðtÞþ b ðtÞ ’ 6 B i a j C7 Hubbel s (2001) neutral theory of biodiversity is the 6 B j6¼i C7 biðÞ¼t þ 1 biðÞt exp6riB1 C7 ð8:1Þ extreme case of an equalizing force where demo- 4 @ Ki A5 graphic parameters are assumed to be identical among species such that interacting with another species has the same per capita impact as interacting Ki is the equilibrium biomass of i in the absence of with a congener. Equalizing mechanisms are not competitors, ri is the intrinsic rate of increase in mathematically stable and cannot allow long-term biomass, and a is the ratio of inter- to intra-specific coexistence. Rather, equalizing mechanisms only interaction. If species have similar carrying capaci- serve to slow the inevitable outcome of species ties and symmetric interactions, then all species interactions. Thus, long-term coexistence requires have the same biomass at equilibrium, b(1), and some type of stabilizing force that involves niche for any local community differentiation in space or time. Regardless of whe- ther occurs through partitioning bð1Þ þ aðS 1ÞbðaÞ¼K ð8:2Þ of limited resources, shared predators, or some other dimension of a species niche, stabilizing forces all From this, the total biomass of the community is share the feature that they reduce interspecificrela- tive to intraspecific interactions, leading to a per SK Bð1Þ ¼ ð8:3Þ capita growth advantage of a species when rare. 1 þ aðS 1Þ The literature is ripe with models that examine how reductions in interspecific relative to intra- For the extreme cases of a ¼ 1ora ¼ 0, eq. 3 specific interactions regulate the impacts of species reduces to B ¼ K and B ¼ SK, respectively, which diversity on the production of single trophic-level shows that community biomass is independent of, TOWARDS A FOOD WEB PERSPECTIVE ON BIODIVERSITY AND ECOSYSTEM FUNCTIONING 115 or a linear function of richness (Fig. 8.4). For all resist or recover from consumption, this can moderate other scenarios where 0 < a < 1, community bio- coexistence among prey (Holt et al. 1994) and dictate mass is a positive but decelerating function of whether prey biomass increases or decreases with species richness. Importantly, the curvilinearity of diversity (Holt and Loreau 2002, Thebault and Loreau this function has nothing to do with how ‘unique’ 2003). Similarly, the tradeoff between the degree of or ‘redundant’ species are. Rather, the decelerating resource specialization and assimilation efficiency of relationship is an inevitable consequence of pack- consumers has important implications for the BEF ing more species into a finite niche axis. Even relationship. The diversity of consumers that pay no when all species are specialists with a unique cost to generalism, i.e. that do not trade off their niche, the contributions by any single species to ability to consume a wide diversity of resources resource capture and biomass production must against their efficiency at consuming each of these decline as a function of richness (i.e. b _ 1/S,(Eqn resources, typically has a strong destabilizing effect 8.2), causing each increase in diversity to contrib- on both population- and ecosystem-level fluctuations, ute smaller increments to resource capture and whereas species diversity has a stabilizing effect on biomass. ecosystem-level fluctuations when consumers do Equation (8.3) predicts a rather straightforward set have such tradeoffs (Thebault and Loreau 2005). of relationships between species diversity and com- So are the consequences of extinction the same in munity biomass for any trophic group that is sup- single versus multi-trophic systems? Theory pre- ported by a non-dynamic resource (e.g. plants dicts that the answer entirely depends on the form assimilating inorganic resources, detritivores feeding of tradeoffs that mediate the coexistence of both on dead organic matter, etc.). One of the key questions consumers and their resources, and whether or not as we extend BEF theory to multi-trophic systems is resources exhibit density dependent dynamics and whether this same set of simple relationships holds by consumers. What we need now true for systems where the resources are themselves are innovative experiments that manipulate the dynamic. Interestingly, several authors have ana- strength of consumer–resource interactions and/or lyzed Lotka–Volterra models for both dynamic and the existence of tradeoffs that are presumed to non-dynamic resources and found that the effects of underlie diversity effects in multi-trophic systems. species diversity on community biomass are often Although such innovative experiments will no qualitatively similar between one and two trophic- doubt be challenging, they have the potential to level systems (Thebault and Loreau 2003, Thebault yield some of the most important new insights into and Loreau 2006, Ives et al. 2005, Fox 2004b). There the functioning of food webs. seems to be just two general instances where new behaviors emerge in a multi-trophic system. The first 8.4.2 Realistic scenarios of extinction occurs when dynamic resources, which have the potential to be overexploited in a multi-trophic sys- It is well established that species extinction is a non- tem, are brought to extinction by their consumers. random processes. Throughout both geological and Overexploitation or extinction of resources by a modern time, certain biological traits such as dis- diverse group of generalist consumers can yield persal ability, generation time, body size, geo- humped-shaped diversity–biomass relationships in graphic range, and local density have proven to be both predators and prey, which is a BEF relationship correlated with extinction risk (McKinney 1997, that is not found in single trophic-level systems Lawton and May 1995, Purvis et al. 2000a). Trophic (Thebault and Loreau 2003, Thebault and Loreau position also appears to be correlated with extinc- 2006, Ives et al. 2005). Second, there are certain types of tion risk. In marine systems, extinction of fish spe- life-history tradeoffs that can alter the shape and cies generally proceeds from the top of food webs magnitude of a diversity–biomass relationship (The- downward (Pauly et al. 1998), which is partly due bault and Loreau 2003, Thebault and Loreau 2006). to human preferences for large-bodied fish, and For example, when resource species exhibit a tradeoff partly because such fish have low resilience due to between their competitive abilities and their ability to late maturity and slow growth (Myers and 116 BIODIVERSITY, ECOSYSTEM FUNCTIONING, AND HUMAN WELLBEING

1025 Sm 1022 < 1 generation > 1 generation

Cl 1019

1016 Bu 1013

10 Un or L)/mean organism size (g) or L)/mean organism 10

2 Tr Spatial scale 7 10 Mo Wo 104

101

10–2 Size of study system (m

10–5 10–3 10–2 10–1 10–0 101 102 103 104 105 106 Time-scale Duration of experiment (d)/mean generation time (d)

Figure 8.5 The spatial and temporal scale of biodiversity-ecosystem functioning experiments. The experimental duration (in days) and spatial scale (in m2 or L) of experiments reviewed by Cardinale et al. (2006) were standardized to the mean generation time and body sizes of the focal organisms. Data were divided into four trophic groups: Plants ¼ green circles, Herbivores ¼ blue triangles, Predators ¼ red squares, Detritivores ¼ brown diamonds. The scale of each individual study is given by smaller symbols while the medians for each trophic group are shown as larger symbols. The box denoted by the dashed line gives the 10th and 90th percentiles for the scale of all experiments. For comparative purposes we show the scale of several natural extinctions: Wo ¼ from Yellowstone National Park, USA; Mo ¼ Moa from New Zealand; Tr ¼ Trout from Lake Superior, USA; Un ¼ Unionid mussels from the lower Mississippi River, USA; Bu ¼ Various species of butterflies in Europe; Cl ¼ Loss of certain cladoceran from Lake Superior, USA; Sm ¼ Global eradication of the small pox virus.

2005). In terrestrial systems, studies similarly report random extinctions). Predators may have high higher extinction probabilities for predators than functional importance in food webs, first because of their prey (Kruess and Tscharntke 1994, Didham the strength of top-down processes in food webs et al. 1998b). (Duffy 2003), and second because predators may Non-random patterns of extinction can affect have traits that are additionally correlated with high diversity–function relationships in at least two ways: functional impact (e.g. body size – Solan et al. 2004). via the functional traits lost, and via changes in Following extinction of a species, diversity–function community interactions. Initially, ecosystem function relationships are additionally influenced by the res- may be most affected by the functional traits of the ponse of the surviving species to loss of a community species that preferentially go extinct (Srivastava and member. Gross and Cardinale (2005) showed that Vellend 2005, Lavorel and Garnier 2002). Positive the effect of species interactions amongst survivors covariance between extinction risk and the magni- depends critically on the mechanisms that underlie tude (Gross and Cardinale 2005) or uniqueness diversity–function relationships: niche partitioning, (Petchey and Gaston 2002b) of a species functional facilitation or the sampling effect each make very effects can exacerbate the impacts of species loss on different predictions about how biased extinction ecosystem function (i.e. diversity–function effects scenarios differ from random extinction scenarios. are initially stronger for realistic extinctions than In food web simulations, Ives and Cardinale (2004) TOWARDS A FOOD WEB PERSPECTIVE ON BIODIVERSITY AND ECOSYSTEM FUNCTIONING 117 showed that the coupling of directional extinction persal as a process affecting species coexistence with species interactions can lead to unexpected becomes particularly prominent at higher trophic changes in the functional importance of species. levels where organisms are typically more mobile (at Although it is clear that non-random patterns of least, on the shorter time-scales of most experiments) extinction can have very different implications than and, therefore, have the ability to integrate informa- the random extinctions that commonly simulated in tion across a landscape and aggregate in response to experiments, our ability to predict the functional the density of their prey. This is important because changesthatstemfromnon-randomextinction– dispersal and aggregation across spatially distinct particularly the top-down effects of species loss on patches or boundaries can translate into ecosystem function – is still in its infancy. After our various forms of niche partitioning that stabilize need to characterize interaction strengths and inter- competitive interactions and consumer–resource specific tradeoffs (Section 8.4.1), our single biggest gap dynamics (Armstrong 1976, McCann et al. 2005). As it of knowledge stems from a lack of information about modifies coexistence, dispersal across patches or levels of covariance between extinction risk and spe- habitat boundaries can also qualitatively alter the BEF cies-specific impacts on rates of ecological processes at relationship (Mouquet et al. 2002). various trophic levels. Although most of the work that has examined how dispersal affects BEF relationships has focused on single trophic level systems, it is useful to quickly 8.4.3 Environmental heterogeneity, patch review here and then consider how these predictions dynamics, and scale might be extended to systems with dynamics resources. A wide variety of ecological models have The typical biodiversity experiment performed to date highlighted the important role that dispersal plays in has taken place in experimental units slightly larger maintaining the diversity of communities (e.g. Island than a five-gallon bucket, and has run for less than one Theory – MacArthur and Wilson 1967, generation of the focal organisms (Fig. 8.5). While ‘mass’ effects – Shmida and Wilson 1985, ‘rescue’ there are noteworthy exceptions (Tilman et al. 2001, effects – Brown and Kodricbrown 1977). Historically, Hector et al. 1999), it seems safe to say that most of our models of dispersal have been phenomenological, inferences about biodiversity stem from experiments meaning they did not explain the existence of diver- performed at spatial scales much smaller, and tem- sity based on first principles. Instead, these models poral scales much shorter than those at which species assumed there was some ‘magical’ pool of species that extinctions actually matter (also see Naeem 2001a for a coexisted at large scales via some unknown mecha- more complete review). Overcoming this mismatch in nism(s), and these species generated propagules that scale is a daunting task, and the difficulties of per- could subsidize local populations. The of forming large-scale, long-term experiments are why meta-community theory (Leibold et al. 2004) repre- ecologists use simplified model systems in the first sented a major advance because these models place (Srivastava et al. 2004). Nevertheless, ecologists acknowledged that everything in a propagule pool have begun to make progress on these issues by must ultimately come from the collection of patches incorporating the important ecological factors that or habitats that span a species range. Based on first co-vary with scale into their experimental designs principles, meta-community models predict both the (Cardinale and Palmer 2002, Dimitrakopoulos and causes and consequences of diversity at ‘local’ Schmid 2004, Mulder et al. 2001) and accounting for (organisms interacting as communities within pat- them in meta-analyses of experiments performed at ches) and ‘regional’ scales (patches of communities different scales (Cardinale et al. 2007). connected by dispersal). The issue of scale is by no means unique to BEF One common form of meta-community models research, nor is it specific to multi-trophic systems. assumes that species coexist through tradeoffs in There are, however, certain characteristics of multi- their abilities to compete in patches that have dif- trophic systems that make it especially important that fering types or supply rates of resources (i.e. what we deal more directly with the issue. Namely, dis- Leibold et al. 2004 call ‘species-sorting’ models). 118 BIODIVERSITY, ECOSYSTEM FUNCTIONING, AND HUMAN WELLBEING

These models predict that at the scale of any local sively.). The general idea of the spatial insurance community, increasing the number of species in the hypothesis is that while one species may be sufficient meta-community serves only to ensure that species to maximize production in any local community, the best adapted to a given patch will colonize and maximization of across all patches in any dominate that patch. This is the typical ‘selection heterogeneous landscape requires that a diversity of effect’ of diversity (Loreau and Hector 2001, Huston species exhibit niche differences at a regional scale. 1997), which has been formalized as follows: Meta-community models like that used to generate assume that species can be ranked by their carrying the spatial insurance hypothesis are important capacities such that K(m) represents the species because they serve as a springboard from which we having the highest in any single can address more pressing issues within the field of patch, K(m-1) is the next highest, and so on. If com- BEF research. From the perspective of basic theory, petition among species is strong (a ¼ 1 in Eqn 8.1), we need to extend meta-community models to con- only one species will be present in a patch at sider how species diversity impacts the production of equilibrium, and the biomass in a patch will be community biomass when consumers and their resources both move across a spatially heterogeneous N N N Bð1Þ ¼ col K þ 1 col col K landscape. We need to know what happens to BEF ðmÞ 1 ðm1Þ g g g ð8:4Þ relationships when (1) resources have a spatial refuge N N N þ 1 col 1 col col K þ ::: from their consumers, (2) consumers and resources 1 2 ðm2Þ g g g disperse at similar versus different rates, or (3) species exhibit spatially mediated tradeoffs, such as in their Equation (8.4) says that the amount of biomass dispersal versus competitive abilities, or dispersal produced in a patch at equilibrium is proportional versus ability to resist consumption. At the same time, to the probability, N /g, that the species with the col we need experiments that explicitly mimic the highest carrying capacity, K , will colonize the (m) assumptions of different meta-community models, patch. If a patch is not colonized by the most and then examine how diversity impacts the pro- productive species, then the probability that the duction of local and regional biomass for various second most productive species, K(m–1), will colo- N N mechanisms that allow consumer–resource coexis- nize and dominate the patch is 1 col col . g g1 tence. These advances are essential if we expect to Note that as the number of species colonizing a patch increases, the probability that a patch predict the ecological consequences of extinction becomes dominated by the most productive spe- from real food webs where the norm is that species cies in the regional approaches unity. move across habitat boundaries and make choices However, one key point is that for the selection about where to spend their time in order to maximize effect to operate in the first place, species diversity fitness. must first exist in the regional colonist pool (i.e. g must exist at the scale of a meta-community). But in order for diversity to be maintained in the 8.4.4 Socio-economic impacts of food web regional colonist pool, species must exhibit some diversity form of tradeoff that ensures they use resources in After several decades of research, it has become ways that are complementary across patches. This apparent that loss of diversity from an ecosystem can suggests that the same mechanisms that ensure complementary use of resources across patches in have impacts on ecological processes that rival, if not a region also produce species-specific selection exceed, many other forms of environmental change. effectsatthescaleofalocalcommunity(Cardinale Ecologists are now in a position to estimate the et al. 2004). number of species required to maximize the removal Loreau et al. (2003) similarly showed that coexis- of greenhouses gasses like CO2 from the atmosphere, tence of species at a regional scale could maximize remove nutrient pollutants from streams and lakes biological production at a local scale, and called this that serve as drinking water, or to produce crops and the ‘spatial insurance’ hypothesis of diversity (also see fisheries. Indeed, it is now possible to make reason- Chapter 10, where Gonzalez treats the issue exten- ably educated estimates of how diversity loss TOWARDS A FOOD WEB PERSPECTIVE ON BIODIVERSITY AND ECOSYSTEM FUNCTIONING 119

Box 8.1 Socioeconomic impacts of predator diversity

One of the primary services that ecosystems provide to aphids, the parasitoid wasp became more efficient at society is the biological control of insect pests. This attacking the pea aphid. As a result, when all three service is estimated to be worth US$400 billion per year enemies were together they reduced aphid populations to globally (Costanza et al. 1997). Although it has long been one-half of that achieved by any enemy species alone. assumedthateffectivepestmanagementrequiresa This translated to a 51 per cent increase in the yield of diversity of predators, parasites, and pathogens (collectively alfalfa. Alfalfa is the fourth most widely grown crop in the called ‘natural enemies’), experiments designed to USA with an estimated annual value of US$11.7 billion explicitly test this hypothesis have only recently begun. (source: US Department of Agriculture). In 2003 when Two case studies highlight the range of results observed this study was performed, alfalfa was selling for $150 per thus far. acre. The state of Wisconsin dedicates 3.5 million acres to the production of alfalfa. Assuming the results of this AQ: Please experiment can be generalized to Wisconsin, the provide economic benefit of predator diversity would be roughly US caption for this figure. $525 million during a single harvesting cycle. In a typical year in the midwestern USA, alfalfa is harvested 3· per summer.

FPO Case study 2: Predator diversity increases pest populations

In a second field experiment, Cardinale et al. (2006) manipulated the diversity of a different group of aphid predators, this time focusing on three species of ladybeetles that are all generalist predators. When the ladybeetles were placed together in field enclosures, they tended to compete with each other in a way that reduced their individual ability to capture prey. As a result, Case study 1: Predator diversity decreases pest more diverse predator assemblages were roughly 60 per fi populations cent less ef cient at controlling aphid populations than expected based on how each ladybeetle performed In a field experiment performed in Wisconsin, USA, when alone. In this case, the antagonistic interactions Cardinale et al. (2003) manipulated the richness of three among the predators led to a 17 per cent decrease in natural enemies of aphids (pea and cowpea) that are alfalfa yield. This result emphasizes that predator species herbivorous pests of alfalfa. Two of the enemies – a can interact in ways that may have economic costs. A key ladybeetle and an assassin bug – were generalist predators challenge for ecologists is to determine the frequency of that fed on both aphid species. The third was a specialist positive and negative interactions among predators that parasitoid wasp that attacks only pea aphids. They found might help us evaluate the costs versus benefits of that as generalist predators reduced the density of both biodiversity.

translates into societally meaningful units – whether predators, parasitoids, and pathogens that be in dollars, health risks, carbon credits, or can be important promoters of top-down control in otherwise. terrestrial food webs, helping to keep pests below The socio-economic implications of biodiversity economically damaging levels. This natural bio- are perhaps most obvious from studies of higher- logical control of pests represents a valuable eco- trophic levels, including those of pollinators, and system service that is essential to sustainable of natural enemies that control pest populations. production of food and fibre. Recent economic 120 BIODIVERSITY, ECOSYSTEM FUNCTIONING, AND HUMAN WELLBEING valuation of the services provided by insects sug- (Ricketts et al. 2004). At a more regional level, Losey gests the value of biological control of native pests and Vaughan (2006) calculate that native pollinators by natural enemies is $4.49 billion per year in the (mostly bees) may be responsible for > $3 billion of USA alone (Losey and Vaughan 2006), and > US fruit and vegetables produced in the USA. $400 billion per year at a global scale (Costanza Although often less direct, changes in biodiver- et al. 1997). While classical biological control tends sity and associated trophic structure have major to focus on the contribution of individual species of implications for issues such as disease risk, with natural enemies, a growing number of studies sug- associated impacts on economics and human well gest that the efficiency of biocontrol is often a func- being. For example, top predators are often the first tion of non-additive interactions among multiple species to disappear as habitat is destroyed and predators, parasitoids, and pathogens (Rosenheim fragmented. As elaborated in Chapter 15, when 2007, Losey and Denno 1998, Snyder and Ives 2003, predators are lost to ecosystems, their prey may Snyder et al. 2006, Finke and Denno 2005, Cardinale increase in abundance, leading to increased trans- et al. 2003, Cardinale et al. 2006b). Although it is not mission efficiency of zoonotic diseases such as yet clear whether these interactions among enemies Lyme disease (Ostfeld and Holt 2004, Dobson et al. generally increase or decrease prey populations, it is 2006). While quantifying the benefit of biodiversity clear that the economic impacts of natural enemy in terms of disease regulation and infected cases diversity can be substantial (Box 8.1). averted is clearly complex, many diseases such as Crop pollination is another malaria, tick-borne encephalitis, and West Nile centred on interactions across trophic levels. The fever have been shown to increase as biodiversity global value of pollination services have been esti- falls (Dobson et al. 2006, and Chapter 15). mated at US$117 billion per year (Costanza et al. 1997) and in a recent review, Klein et al. (2007) 8.5 Summary concluded that fruit, vegetable, or seed biodiversity (i.e. richness, abundance, and distribution of mul- The emerging paradigm of Biodiversity Effects on tiple species of pollinators) in delivering this eco- Ecosystem Functioning has shown great potential system service are often poorly quantified. Similar to augment ecology’s historical focus on the causes to evaluations of classical biocontrol, where the of biodiversity with a much more contemporary focus is on the action of one or few natural enemies understanding of its ecological consequences. Even rather than diversity per se, many of the economic so, BEF studies have, thus far, been limited to valuations of pollination services consider the con- highly simplified ‘model’ communities that are tribution of honey bees alone. Klein et al. (2007) nowhere near the trophic complexity of real com- report case studies for nine crops on four continents munities. To overcome this limitation, it is now implicating a diversity of pollinators and revealing imperative that ecologists begin to merge the BEF that agricultural intensification jeopardizes wild bee paradigm with more classic ideas in food web communities and their stabilizing effect on polli- ecology that detail how interactions among trophic nation services at the landscape scale. At the indi- levels that play out in space and time can constrain vidual farm level, such natural pollination services fluxes of energy and matter. Most hypotheses about can contribute significantly to annual income; a the functional role of diversity within and across study from a coffee plantation in Costa Rica, for trophic levels are in their infancy, and they repre- example, indicated native bee species account for sent a rich opportunity for new work during the $62,000, or 7 per cent of the farm’s annual income second generation of BEF experiments.